Throughout this application, certain publications are referenced. Full citations for these publications, as well as additional related references, may be found immediately preceding the claims. The disclosures of these publications are hereby incorporated by reference into this application in order to more fully describe the state of the art as of the date of the invention described and claimed herein.
This application relates to musculoskeletal tissue engineering. For example, a scaffold apparatus is discussed below which can serve as a functional interface between multiple tissue types. Methods for preparing a multi-phase scaffold are also discussed. Some exemplary embodiments which include a soft tissue-bone interface are discussed.
As examples of soft tissue-bone interface, the human rotator cuff and anterior cruciate ligament (ACL) are described below. The rotator cuff, the ACL and the ACL-bone interface are used in the following discussion as examples to aid in understanding the description of the methods and apparatuses of this application. This discussion, however, is not intended to, and should not be construed to, limit the claims of this application.
The Rotator Cuff
The rotator cuff consists of a group of four muscles and tendons, including the supraspinatus, infraspinatus, teres minor, and subscapularis, which function in synchrony to stabilize the glenohumeral joint as well as to actively control shoulder kinematics. The supraspinatus tendon inserts into the humeral head via a direct insertion exhibiting region-dependent matrix heterogeneity and mineral content.
Four distinct yet continuous tissue regions are observed at the tendon-bone junction (FIG. 9A): tendon proper, non-mineralized fibrocartilage, mineralized fibrocartilage and bone (Benjamin, 1986; Benjamin, 2002; Woo, 1988). The tendon proper consists of fibroblasts found between aligned collagen fibers in a matrix rich in collagen I, with small amounts of collagen III and proteoglycans (Blevins, 1997). The non-mineralized fibrocartilage region is composed of fibrochondrocytes in a matrix of collagen I, II, and III with fibers oriented perpendicular to the calcified interface region (Kumagai, 1994). The mineralized fibrocartilage region consists of hypertrophic fibrochondrocytes within a matrix of collagen I and II (Kumagai, 1994) as well as collagen X (Thomopoulos, 2003). The last region of the insertion site is bone which consists of osteoblasts, osteoclasts, and osteocytes in a mineralized matrix rich in type I collagen. This controlled matrix heterogeneity exhibited by the tendon-bone interface serves to minimize stress concentrations and to mediate load transfer between two distinct tissue types (Thomopoulos, 2003; woo, 1988). Due to its functional significance, interface regeneration is a pre-requisite for biological fixation.
Rotator cuff tears are among the most common injuries afflicting the shoulder, with greater than 75,000 repair procedures performed annually in the United States alone (Vitale, 2007). Clinical intervention is required because injuries to the rotator cuff do not heal, largely due to the complex anatomy noted above and the extended range of motion of the shoulder joint, as well as the relative weakening and hypovascularization of the cuff tendons (Codman, 1934; Yamanaka, 1994; Dejardin, 2001). Moreover, chronic degeneration increases both the frequency and size of cuff tears with age (Tempelhof, 1999) and is considered the main contributing factor in the pathogenesis of rotator cuff tendon tears (Dejardin, 2001; Soslowsky, 2000). Early primary anatomic repair followed by carefully controlled rehabilitation is currently the treatment of choice for symptomatic rotator cuff tears (Dejardin, 2001).
Rotator cuff repair has evolved from traditional open repair to “mini-open” to primarily arthroscopic (Gartsman, 2001; Galatz, 2004; Willaims, 2004; Mazzocca, 2005; Cole 2007). This transition has occurred due to advances in surgical techniques and fixation methods, with the current technique being a double row “suture-bridge” technique which simulates the compression afforded by transosseous sutures previously used in open and mini-open repairs (Park, 2007). These methods have been shown to improve mechanical strength and graft stability (Park, 2005). The focus in the field now centers on how to address the challenge of achieving functional rotator cuff healing and/or augmentation, which is essential for long term clinical success.
Currently, a significant demand exists for a functional tendon grafting system which can augment and promote rotator cuff healing post surgical repair, due to the relatively high failure rates associated with current repair procedures as well as the clinical need to treat large tears and chronic degeneration of the rotator cuff tendons. For example, failure rates as high as 90% have been reported after primary repair of chronic rotator cuff injuries (Galatz, 2004), generally attributed to factors such as osteoporotic bone, degenerative and poorly vascularized tendons, severe tendon weakening, muscle atrophy, and size of the original defect (Gazielly, 1994; Gartsman, 1997; Mansat, 1997; Rokito, 1996; Romeo, 1999). Moreover, the primary repair of chronic degenerative cuff injuries often results in excessive tension on the cuff tissues and at the repair site (Dejardin, 2001; DeOrio, 1984; Gerber, 1999). To improve healing, synthetic grafts (Post, 1985; Ozaki, 1986) have been designed to reconstruct large rotator cuff defects. See also, e.g., U.S. Pat. No. 7,112,417. However, these devices are suboptimal due to concerns of biocompatibility as well as their inability to meet the functional demand of the native tendon.
Recently, biological matrices such as acellularized allogeneic and xenogeneic extracellular matrix scaffolds have emerged as promising grafts for rotator cuff repair and augmentation (Dejardin, 2001; Schlpegel, 2006). Both collagen-rich dermis and small intestinal submucosa (SIS) (Badylak, 2002) have been marketed commercially as graft patches for reinforcing soft tissue repair following rotator cuff surgery (Derwin, 2006; Lannotti, 2006). See also, e.g., U.S. Pat. Nos. 6,638,312 and 7,160,333. SIS is particularly attractive as it exhibits a biomimetic, collagen nanofiber-based architecture and alignment, thus it can be readily remodeled by host cells while encouraging angiogenesis and neo-collagen production (Badylak, 2002).
Highly promising results have been reported for SIS in several animal models (Dejardin, 2001; Schlegel, 2006), but unfortunately suboptimal outcomes were observed in human trials (Lannotti, 2006; Sclamberg, 2004), in which augmentation with SIS did not improve the rate of tendon healing or clinical outcome scores. Similar outcomes have been reported for other biological grafts used in rotator cuff repair (Sclamberg, 2004; Coons, 2006).
The suboptimal results of biologically-derived grafts may be attributed to mismatch in mechanical properties and the rapid matrix remodeling experienced in the physiologically demanding and often diseased shoulder joint. Utilizing a canine model, Derwin et al. (Derwin, 2006) performed a systematic comparison of the biomechanical properties of commercially available extracellular matrices for rotator cuff augmentation. A mismatch in mechanical properties with the canine infraspinatus tendon was observed for all types of extracellular matrix tested.
Moreover, it has been reported that the mechanical properties of SIS decreased as resorption occurred prematurely (Derwin, 2006). Thus the mismatch in the kinetics of graft remodeling and neo-collagen formation compromised the clinical outcome. Therefore, the debilitating effect of rotator cuff tears coupled with the high incidence of failure associated with existing graft choices emphasize the clinical need for functional rotator graft augmentation solutions.
The Anterior Cruciate Ligament (ACL)
The ACL consists of a band of regularly oriented, dense connective tissue that spans the junction between the femur and tibia. It participates in knee motion control and acts as a joint stabilizer, serving as the primary restraint to anterior tibial translation. The natural ACL-bone interface consists of three regions: ligament, fibrocartilage (non-mineralized and mineralized) and bone. The natural ligament to bone interface is arranged linearly from ligament to fibrocartilage and to bone. The transition results in varying cellular, chemical, and mechanical properties across the interface, and acts to minimize stress concentrations from soft tissue to bone.
The ACL is the most often injured ligament of the knee. Due to its inherently poor healing potential and limited vascularization, ACL ruptures do not heal effectively upon injury, and surgical intervention is typically needed to restore normal function to the knee.
Clinically, autogenous grafts based on either bone-patellar tendon-bone (BPTB) or hamstring-tendon (HST) grafts are often a preferred grafting system for ACL reconstruction, primarily due to a lack of alternative grafting solutions. Current ACL grafts are limited by donor site morbidity, tendonitis and arthritis. Synthetic grafts may exhibit good short term results but encounter clinical failure in long-term follow-ups, since they are unable to duplicate the mechanical strength and structural properties of human ACL tissue. ACL tears and ruptures are currently commonly repaired using semitendinosus grafts. Although semitendinosus autografts are superior, they often fail at the insertion site between the graft and the bone tunnel. One of the major causes of failure in this type of reconstruction grafts is its inability to regenerate the soft-tissue to bone interface.
Despite their distinct advantages over synthetic substitutes, autogenous grafts have a relatively high failure rate. A primary cause for the high failure rate is the lack of consistent graft integration with the subchondral bone within bone tunnels. The site of graft contact in femoral or tibial tunnels represents the weakest point mechanically in the early post-operative healing period. Therefore, success of ACL reconstructive surgery depends heavily on the extent of graft integration with bone.
ACL reconstruction based on autografts often results in loss of functional strength from an initial implantation time, followed by a gradual increase in strength that does not typically reach the original magnitude. Despite its clinical success, long term performance of autogenous ligament substitutes is dependent on a variety of factors, including structural and material properties of the graft, initial graft tension, intrarticular position of the graft, as well as fixation of the graft. These grafts typically do not achieve normal restoration of ACL morphology and knee stability.
There is often a lack of graft integration with host tissue, in particular at bony tunnels, which contributes to suboptimal clinical outcome of these grafts. The fixation sites at the tibial and femoral tunnels, instead of the isolated strength of the graft material, have been identified as mechanically weak points in the reconstructed ACL. Poor graft integration may lead to enlargement of the bone tunnels, and in turn may compromise the long term stability of the graft.
Increased emphasis has been placed on graft fixation, as post surgery rehabilitation protocols require the immediate ability to exercise full range of motion, reestablish neuromuscular function and weight bearing. During ACL reconstruction, the bone-patellar tendon-bone or hamstring-tendon graft is fixed into the tibial and femoral tunnels using a variety of fixation techniques. Fixation devices include, for example, staples, screw and washer, press fit EndoButton® devices, and interference screws. In many instances, EndoButton® devices or Mitek® Anchor devices are utilized for fixation of femoral insertions. Staples, interference screws, or interference screws combined with washers can be used to fix the graft to the tibial region.
Recently, interference screws have emerged as a standard device for graft fixation. The interference screw, about 9 mm in diameter and at least 20 mm in length, is used routinely to secure tendon to bone and bone to bone in ligament reconstruction. Surgically, the knee is flexed and the screw is inserted from the para-patellar incision into the tibial socket, and the tibial screw is inserted just underneath the joint surface. After tension is applied to the femoral graft and the knee is fully flexed, the femoral tunnel screw is inserted. This procedure has been reported to result in stiffness and fixation strength levels which are adequate for daily activities and progressive rehabilitation programs.
While the use of interference screws have improved the fixation of ACL grafts, mechanical considerations and biomaterial-related issues associated with existing screw systems have limited the long term functionality of the ligament substitutes. Screw-related laceration of either the ligament substitute or bone plug suture has been reported. In some cases, tibial screw removal was necessary to reduce the pain suffered by the patient. Stress relaxation, distortion of magnetic resonance imaging, and corrosion of metallic screws have provided motivation for development of biodegradable screws based on poly-α-hydroxy acids. While lower incidence of graft laceration was reported for biodegradable screws, the highest interference fixation strength of the grafts to bone is reported to be 475 N, which is significantly lower than the attachment strength of ACL to bone. When tendon-to-bone fixation with polylactic acid-based interference screws was examined in a sheep model, intraligamentous failure was reported by 6 weeks. In addition, fixation strength is dependent on quality of bone (mineral density) and bone compression.
Two insertion zones can be found in the ACL, one at the femoral end and another located at the tibial attachment site. The ACL can attach to mineralized tissue through insertion of collagen fibrils, and there exists a gradual transition from soft tissue to bone. The femoral attachment area in the human ACL was measured to be 113±27 mm2 and 136±33 mm2 for the tibia insertion. With the exception of the mode of collagen insertion into the subchondral bone, the transition from ACL to bone is histologically similar for the femoral and tibial insertion sites.
The insertion site is comprised of four different zones: ligament, non-mineralized fibrocartilage, mineralized fibrocartilage, and bone. The first zone, which is the ligament proper, is composed of solitary, spindle-shaped fibroblasts aligned in rows, and embedded in parallel collagen fibril bundles of 70-150 μm in diameter. Primarily type I collagen makes up the extracellular matrix, and type III collagen, which are small reticular fibers, are located between the collagen I fibril bundles. The second zone, which is fibro-cartilaginous in nature, is composed of ovoid-shaped chondrocyte-like cells. The cells do not lie solitarily, but are aligned in rows of 3-15 cells per row. Collagen fibril bundles are not strictly parallel and much larger than those found in zone 1. Type II collagen is now found within the pericellular matrix of the chondrocytes, with the matrix still made up predominantly of type I collagen. This zone is primarily avascular, and the primary sulfated proteoglycan is aggrecan. The next zone is mineralized fibrocartilage. In this zone, chondrocytes appear more circular and hypertrophic, surrounded by larger pericellular matrix distal from the ACL. Type X collagen, a specific marker for hypertrophic chondrocytes and subsequent mineralization, is detected and found only within this zone. The interface between mineralized fibrocartilage and subjacent bone is characterized by deep inter-digitations. Increasing number of deep inter-digitations is positively correlated to increased resistance to shear and tensile forces during development of rabbit ligament insertions. The last zone is the subchondral bone and the cells present are osteoblasts, osteocytes and osteoclasts. The predominant collagen is type I and fibrocartilage-specific markers such as type II collagen are no longer present.
For bone-patellar tendon-bone grafts, bone-to-bone integration with the aid of interference screws is the primary mechanism facilitating graft fixation. Several groups have examined the process of tendon-to-bone healing.
Blickenstaff et al. (1997) evaluated the histological and biomechanical changes during the healing of a semitendinosus autograft for ACL reconstruction in a rabbit model. Graft integration occurred by the formation of an indirect tendon insertion to bone at 26 weeks. However, large differences in graft strength and stiffness remained between the normal semitendinosus tendon and anterior cruciate ligament after 52 weeks of implantation.
In a similar model, Grana et al. (1994) reported that graft integration within the bone tunnel occurs by an intertwining of graft and connective tissue and anchoring of connective tissue to bone by collagenous fibers and bone formation in the tunnels. The collagenous fibers have the appearance of Sharpey's fibers seen in an indirect tendon insertion.
Rodeo et al. (1993) examined tendon-to-bone healing in a canine model by transplanting digital extensor tendon into a bone tunnel within the proximal tibial metaphysis. A layer of cellular, fibrous tissue was found between the tendon and bone, and this fibrous layer matured and reorganized during the healing process. As the tendon integrated with bone through Sharpey's fibers, the strength of the interface increased between the second and the twelfth week after surgery. The progressive increase in strength was correlated with the degree of bone in growth, mineralization, and maturation of the healing tissue.
In most cases, tendon-to-bone healing with and without interference fixation does not result in the complete re-establishment of the normal transition zones of the native ACL-bone insertions. This inability to fully reproduce these structurally and functionally different regions at the junction between graft and bone is detrimental to the ability of the graft to transmit mechanical stress across the graft proper and leads to sites of stress concentration at the junction between soft tissue and bone.
Zonal variations front soft to hard tissue at the interface facilitate a gradual change in stiffness and can prevent build up of stress concentrations at the attachment sites.
The insertion zone is dominated by non-mineralized and mineralized fibrocartilage, which are tissues adept at transmitting compressive loads. Mechanical factors may be responsible for the development and maintenance of the fibrocartilagenous zone found at many of the interfaces between soft tissue and bone. The fibrocartilage zone with its expected gradual increase in stiffness appears less prone to failure.
Benjamin et al. (1991) suggested that the amount of calcified tissue in the insertion may be positively correlated to the force transmitted across the calcified zone.
Using simple histomorphometry techniques, Gao et al. determined that the thickness of the calcified fibrocartilage zone was 0.22±0.7 mm and that this was not statistically different from the tibial insertion zone. While the ligament proper is primarily subjected to tensile and torsional loads, the load profile and stress distribution at the insertion zone is more complex.
Matyas et al. (1995) combined histomorphometry with a finite element model (FEM) to correlate tissue phenotype with stress state at the medial collateral ligament (MCL) femoral insertion zone. The FEM model predicted that when the MCL is under tension, the MCI, midsubstance is subjected to tension and the highest principal compressive stress is found at the interface between ligament and bone.
Calcium phosphates have been shown to modulate cell morphology, proliferation and differentiation. Calcium ions can serve as a substrate for Ca2+-binding proteins, and modulate the function of cytoskeleton proteins involved in cell shape maintenance.
Gregiore et al. (1987) examined human gingival fibroblasts and osteoblasts and reported that these cells underwent changes in morphology, cellular activity, and proliferation as a function of hydroxyapatite particle sizes. Culture distribution varied from a homogenous confluent monolayer to dense, asymmetric, and multi-layers as particle size varied from less than 5 μg to greater than 50 μg, and proliferation changes correlated with hydroxyapatite particles size.
Cheung et al. (1985) further observed that fibroblast mitosis is stimulated with various types of calcium-containing complexes in a concentration-dependent fashion.
Chondrocytes are also dependent on both calcium and phosphates for their function and matrix mineralization. Wuthier et al. (1993) reported that matrix vesicles in fibrocartilage consist of calcium-acidic phospholipids-phosphate complex, which are formed from actively acquired calcium ions and an elevated cytosolic phosphate concentration.
Phosphate ions have been reported to enhance matrix mineralization without regulation of protein production or cell proliferation, likely because phosphate concentration is often the limiting step in mineralization. It has been demonstrated that human foreskin fibroblasts when grown in micromass cultures and under the stimulation of lactic acid can dedifferentiate into chondrocytes and produce type II collagen.
Cheung at al. (1985) found a direct relationship between β-glycerophosphate concentrations and mineralization by both osteoblasts and fibroblasts. Increased mineralization by ligament fibroblasts is observed with increasing concentration of β-glycerophosphate, a media additive commonly used in osteoblast cultures. These reports strongly suggest the plasticity of the fibroblast response and that the dedifferentiation of ligament fibroblasts is a function of mineral content in vitro.
Progressing through the four different zones which make up the native ACL insertion zone, several cell types are identified: ligament fibroblasts, chondrocytes, hypertrophic chondrocytes and osteoblasts, osteoclasts, and osteocytes. The development of in vitro multi-cell type culture systems facilitates the formation of the transition zones.
No reported studies on either the co-culture of ligament fibroblasts with osteoblasts, nor on the in vitro and in vivo regeneration of the bone-ligament interface are known.
No reported studies which examine the potential of multi-phased scaffolds in facilitating the fixation of ligament or tendon to bone are known. As the interface between graft and bone is the weakest point during the initial healing period, recent research efforts in ACL tissue engineering have concentrated on design of multi-phased scaffolds in order to promote graft integration.
Goulet at al. (2000) developed a bio-engineered ligament model, where ACL fibroblasts were added to the structure and bone plugs were used to anchor the bioengineered tissue. Fibroblasts isolated from human ACL were grown on bovine type I collagen, and the bony plugs were used to promote the anchoring of the implant within the bone tunnels.
Cooper et al. (2000) and Lu et al. (2001) developed a tissue engineered ACL scaffold using biodegradable polymer fibers braided into a 3-D scaffold. This scaffold has been shown to promote the attachment and growth of rabbit ACL cells in vitro and in vivo. However, no multiphased scaffolds for human ligament-to-bone interface are known.